Abstract

Plasma cortisol levels and modified Apache II (Apache IIm-stay) severity of disease scores were determined at weekly intervals in 159 patients who were treated for at least 7 days at the Critical Care Unit of our hospital. The mean (+/- SD) plasma cortisol level (0.60 +/- 0.28 mumol/l) was clearly elevated in these patients. The highest plasma cortisol levels were measured in patients treated with vasoactive drugs (0.76 +/- 0.39 mumol/l). Non-survivors (n = 36) had a significantly higher mean plasma cortisol level and Apache IIm-stay score than survivors (respectively 0.78 +/- 0.40 vs. 0.54 +/- 0.21 mumol/l; p less than 0.0003 and 12.6 +/- 4.8 vs. 7.3 +/- 4.1; p less than 0.0001). A significant correlation was found between the individual weekly plasma cortisol levels and the Apache IIm-stay scores (r = 0.41; p less than 0.0001), especially in the subgroup of patients, who never received glucocorticoids during their stay at the ICU (r = 0.51; p less than 0.0001). During the 14-month study period only two patients showed a clinical picture of adrenocortical insufficiency and a blunted response of cortisol to 0.25 mg synthetic ACTH(1-24). In conclusion, our data suggest that a high plasma cortisol level, like a high Apache IIm-stay score, indicates severity of disease and poor survival in critically ill patients. De novo adrenocortical insufficiency is rare and therefore routine screening of adrenocortical function is superfluous.